Batch editing - In XPATH replace mode don't try to add a node if doesn't exist

[josegar74]

The Batch Edit API uses the following setting to add an element in XPATH replace mode, if the XML node doesn't exist:

This doesn't make much sense, since in replacement mode you want to update an element if it exists, otherwise do nothing.

This pull request updates the code to not use that setting and never try to add the element in XPATH replace mode.

Test case:

1) Create an ISO19139 metadata and add a dataset contact with the name John and role owner 2) In the batch edit configure a XPATH replacement:

• Title: Replace individual name
• Type: Replace the element or value (in all matching parents)
• XPATH: /gmd:MD_Metadata/gmd:identificationInfo/*/gmd:pointOfContact[*/gmd:role/gmd:CI_RoleCode/@codeListValue='owner' and */gmd:individualName/*/text()='John']/*/gmd:individualName/gco:CharacterString
• Value: Mary

The previous change match the element and changes the individual name to Mary

       <gmd:pointOfContact>
          <gmd:CI_ResponsibleParty>
            <gmd:individualName>
              <gco:CharacterString>Mary</gco:CharacterString>
            </gmd:individualName>

3) Edit XPATH expression to match another role: pointOfContact, that doesn't exist in the metadata:

/gmd:MD_Metadata/gmd:identificationInfo/*/gmd:pointOfContact[*/gmd:role/gmd:CI_RoleCode/@codeListValue='pointOfContact' and */gmd:individualName/*/text()='John']/*/gmd:individualName/gco:CharacterString

• Without the fix, produces an invalid ISO19139 element: <gmd:pointOfContact>Mary</gmd:pointOfContact>

• With the fix, the individual name is not updated (stays with the value John) as the XPATH expression does not match.

---

The pull request includes Sonarlint improvements.

---

For existing installations, a workaround to avoid this problem is to disable the settings listed at the beginning, unde r Admin console > Settings > CSW

---

Checklist

• [X] I have read the contribution guidelines
• [X Pull request provided for main branch, backports managed with label
• [X] Good housekeeping of code, cleaning up comments, tests, and documentation
• [X] Clean commit history broken into understandable chucks, avoiding big commits with hundreds of files, cautious of reformatting and whitespace changes
• [X] Clean commit messages, longer verbose messages are encouraged
• [ ] API Changes are identified in commit messages
• [ ] Testing provided for features or enhancements using automatic tests
• [ ] User documentation provided for new features or enhancements in manual
• [ ] Build documentation provided for development instructions in README.md files
• [ ] Library management using pom.xml dependency management. Update build documentation with intended library use and library tutorials or documentation

[josegar74]

@fxprunayre this change breaks some unit tests, like BatchEditsServiceTest.testUpdateRecordAddAttribute:323 expected:<value> but was:<null>

The test is doing this replacement:

https://github.com/geonetwork/core-geonetwork/blob/ee23e5430ef839c90ab6d08890e1176481465741/services/src/test/java/org/fao/geonet/services/metadata/BatchEditsServiceTest.java#L296-L301

I would expect that in the following code the isCreate variable would be true, but that is not the case, so with this change, no longer enters in the condition as createXpathNodeIfNotExist is false and the test fails:

https://github.com/geonetwork/core-geonetwork/blob/ee23e5430ef839c90ab6d08890e1176481465741/core/src/main/java/org/fao/geonet/kernel/EditLib.java#L518-L519

isCreate is false as doesn't check for the tag gn_add, but gn_create.

https://github.com/geonetwork/core-geonetwork/blob/ee23e5430ef839c90ab6d08890e1176481465741/core/src/main/java/org/fao/geonet/kernel/EditLib.java#L505-L507

The difference seems a bit subtle in the documentation, but gn_add is kind of special case of gn_create:

https://github.com/geonetwork/core-geonetwork/blob/ee23e5430ef839c90ab6d08890e1176481465741/core/src/main/java/org/fao/geonet/kernel/EditLib.java#L1821-L1828

I'm afraid changing the code in https://github.com/geonetwork/core-geonetwork/blob/ee23e5430ef839c90ab6d08890e1176481465741/core/src/main/java/org/fao/geonet/kernel/EditLib.java#L505-L507 to check both probably will cause additional unwanted side effects?

[fxprunayre]

gn_add/create don't have effect on attribute which are always text. So a proper fix could be to change AddElemValue to extract that in a specific property... or we can also check if the target is an attribute and in this case enable creation mode.